A Randomized Link Transformer for Diverse Open-Domain Dialogue Generation

Jing Yang Lee, Kong Aik Lee, Woon Seng Gan

Research output: Chapter in book / Conference proceedingConference article published in proceeding or bookAcademic researchpeer-review

1 Citation (Scopus)


A major issue in open-domain dialogue generation is the agent's tendency to generate repetitive and generic responses. The lack in response diversity has been addressed in recent years via the use of latent variable models, such as the Conditional Variational Auto-Encoder (CVAE), which typically involve learning a latent Gaussian distribution over potential response intents. However, due to latent variable collapse, training latent variable dialogue models are notoriously complex, requiring substantial modification to the standard training process and loss function. Other approaches proposed to improve response diversity also largely entail a significant increase in training complexity. Hence, this paper proposes a Randomized Link (RL) Transformer as an alternative to the latent variable models. The RL Transformer does not require any additional enhancements to the training process or loss function. Empirical results show that, when it comes to response diversity, the RL Transformer achieved comparable performance compared to latent variable models.

Original languageEnglish
Title of host publicationACL 2022 - 4th Workshop on NLP for Conversational AI, Proceedings of the Workshop
PublisherAssociation for Computational Linguistics (ACL)
Number of pages11
ISBN (Electronic)9781955917469
Publication statusPublished - May 2022
Externally publishedYes
Event4th Workshop on NLP for Conversational AI, NLP4ConvAI 2022 at ACL 2022 - Dublin, Ireland
Duration: 27 May 2022 → …

Publication series

NameProceedings of the Annual Meeting of the Association for Computational Linguistics
ISSN (Print)0736-587X


Conference4th Workshop on NLP for Conversational AI, NLP4ConvAI 2022 at ACL 2022
Period27/05/22 → …

ASJC Scopus subject areas

  • Computer Science Applications
  • Linguistics and Language
  • Language and Linguistics


Dive into the research topics of 'A Randomized Link Transformer for Diverse Open-Domain Dialogue Generation'. Together they form a unique fingerprint.

Cite this